24-Week study on the use

0300-7995
doi:10.1185/030079908X291967
Current Medical Research and Opinion®
Vol. 24, No. 5, 2008, 1485–1496
© 2008 Informa UK Ltd
All rights reserved: reproduction in whole or part not permitted
ORIGINAL ARTICLE
24-Week study on the use
of collagen hydrolysate as a
dietary supplement in athletes
with activity-related joint pain
Kristine L. Clark a, Wayne Sebastianelli b, Klaus R.
Flechsenhar c, Douglas F. Aukermann b, Felix Meza b,
Roberta L. Millard b, John R. Deitch b, Paul S. Sherbondy b
and Ann Albert a
Department of Nutrition and Sports Nutrition for Athletics, Penn State
University, University Park, PA, USA
a
td
L
Department of Orthopedic Surgery and Rehabilitation,KPenn State
U
University, University Park, PA, USA
a
m Health GmbH,
Department of Research and Development ofrGELITA
o
f
Eberbach, Germany
In
8
Address for correspondence: Kristine Clark, PhD,
00Department of Nutrition and Sports Nutrition for
Athletics, Pennsylvania State University, Room2256 Recreation Hall, University Park, PA 16802, USA.
© klc5@psu.edu
Tel.: +1 814 863 8107; Fax: +1 814 865 1746;
b
c
Key words: Articular cartilage – Athletes – Collagen hydrolysate – Joint pain – Nutritional supplements –
Proteoglycans – Type II collagen
ABSTRACT
Background: Collagen hydrolysate is a nutritional
supplement that has been shown to exert
an anabolic effect on cartilage tissue. Its
administration appears beneficial in patients with
osteoarthritis.
Objective: To investigate the effect of collagen
hydrolysate on activity-related joint pain in
athletes who are physically active and have no
evidence of joint disease.
Design and setting: A prospective, randomized,
placebo-controlled, double-blind study was
conducted at Penn State University in University
Park, Pennsylvania. Parameters including joint
pain, mobility, and inflammation were evaluated
with the use of a visual analogue scale during a
24-week study phase.
Study participants: Between September 2005
and June 2006, 147 subjects who competed on a
varsity team or a club sport were recruited. Data
from 97 of 147 subjects could be statistically
evaluated.
Intervention: One hundred and forty-seven
subjects (72 male, 75 female) were randomly
assigned to two groups: a group (n = 73) receiving
25 mL of a liquid formulation that contained 10 g
of collagen hydrolysate (CH-Alpha)* and a group
(n = 74) receiving a placebo, which consisted of
25 mL of liquid that contained xanthan.
Main outcome measures: The primary efficacy
parameter was the change in the visual analogue
scales from baseline during the study phase
in relation to the parameters referring to pain,
mobility, and inflammation.
Results: When data from all subjects (n =
97) were evaluated, six parameters showed
statistically significant changes with the dietary
supplement collagen hydrolysate (CH) compared
with placebo: joint pain at rest, assessed by
the physician (CH vs. placebo (–1.37 ± 1.78 vs.
–0.90 ± 1.74 ( p = 0.025)) and five parameters
assessed by study participants: joint pain
when walking (–1.11 ± 1.98 vs. –0.46 ± 1.63,
* GELITA Health GmbH, Eberbach, Germany
Paper 4309
1485
p = 0.007), joint pain when standing (–0.97 ±
1.92 vs. –0.43 ± 1.74, p = 0.011), joint pain at
rest (–0.81 ± 1.77 vs. –0.39 ± 1.56, p = 0.039),
joint pain when carrying objects (–1.45 ± 2.11
vs. –0.83 ± 1.71, p = 0.014) and joint pain when
lifting (–1.79 ± 2.11 vs. –1.26 ± 2.09, p = 0.018).
When a subgroup analysis of subjects with knee
arthralgia (n = 63) was performed, the difference
between the effect of collagen hydrolysate vs.
placebo was more pronounced. The parameter
joint pain at rest, assessed by the physician,
had a statistical significance level of p = 0.001
(–1.67 ± 1.89 vs. –0.86 ± 1.77), while the
other five parameters based on the participants’
assessments were also statistically significant:
joint pain when walking ( p = 0.003 ( –1.38 ±
2.12 vs. –0.54 ± 1.65)), joint pain when standing
( p = 0.015 (–1.17 ± 2.06 vs. –0.50 ± 1.68)), joint
Introduction
Collagen hydrolysate (CH-Alpha)* is a dietary supple­
ment that may be beneficial in patients suffering from
degenerative joint disease. Its use in the treatment of
individuals with osteoarthritis (OA) has increasingly
gained support in the medical community and among
consumers. Collagen hydrolysate is made out of
collagenous tissue from porcine sources such as bone,
hide, and hide split. It is a product that is obtained
when these raw materials are subjected to technical
processes including extraction, enzymatic hydrolysis,
purification, concentration, sterilization, and drying.
From the chemical point of view, collagen
hydrolysate consists of proteins that range in size from
0.5 to 13.5 kilodaltons (kDa), with an average size of
3.3 kDa. It is a nongelling substance and can be easily
dissolved in cold water.
In the past two decades, there has been an increased
understanding of the effects of collagen hydrolysate
on joint tissues in various models. Oesser et al., in a
series of preclinical studies, demonstrated that collagen
hydro­lysate passes across the mucosal barrier in the
small bowel as a complete peptide that is no longer
subject to enzymatic cleavage, accumulates in cartilage
tissue, and stimulates production of type II collagen (the
major protein in articular cartilage) and proteoglycans
in the extracellular matrix of cartilage1–3. In addition,
they established a dose–response relationship between
the concentration of collagen hydrolysate in which
chondro­cyte cultures are incubated and the quantity of
type II collagen produced. This finding was confirmed
by Ng et al., who used a three-dimensional approach to
culture chondrocytes4.
In addition to these preclinical studies, open label,
comparative, and prospective, randomized, placebo-
pain at rest with ( p = 0.021 (–1.01 ±1.92 vs.
–0.47 ± 1.63)), joint pain when running a straight
line ( p = 0.027 (–1.50 ± 1.97 vs. –0.80 ± 1.66))
and joint pain when changing direction ( p = 0.026
(–1.87 ± 2.18 vs. –1.20 ± 2.10)).
Conclusion: This was the first clinical trial of
24-weeks duration to show improvement of joint
pain in athletes who were treated with the dietary
supplement collagen hydrolysate. The results of
this study have implications for the use of collagen
hydrolysate to support joint health and possibly
reduce the risk of joint deterioration in a high-risk
group. Despite the study’s size and limitations,
the results suggest that athletes consuming
collagen hydrolysate can reduce parameters (such
as pain) that have a negative impact on athletic
performance. Future studies are needed to support
these findings.
controlled clinical trials and experimental findings have
been published, with several studies providing evidence
of a beneficial effect on measurements of joint health
from the administration of collagen hydrolysate in a
variety of patient populations5–12.
Most of these clinical studies have been conducted
in patients diagnosed with OA5–7,9–11. In the majority of
these studies, investigators were able to demonstrate
that the administration of 7–10 g of collagen hydrolysate
per day for 3 months produced an improvement in
measure­m ents of joint health or function, such as
reduction in pain, 5,6,9 decreased dependency on pain
medications,9 and improvement in leg strength11.
Although these studies demonstrated reduction in
joint pain in patients with OA treated with collagen
hydrolysate, investigators wanted to determine the
benefit of collagen hydrolysate in individuals with
joint pain associated with athletic activities who
had not been diagnosed with a medical disorder. To
address this question, Flechsenhar et al. conducted
an observational study of athletes with joint pain
who did not have a diagnosis of OA at the Olympic
training site (for German Olympic Games qualifiers)
in Essen, Germany. Specifically, the aim of the
study was to determine if treatment with 10 g/day of
collagen hydrolysate would reduce joint pain in these
athletes12.
One hundred athletes with exercise-related joint
pain received 10 g/day of collagen hydrolysate for
12 weeks. Seventy-eight percent of the subjects in this
observational study reported improved symptoms with
reduction of pain on movement in 12 weeks. Similar
percentages of patients reported improvements in pain
while climbing stairs or pain when carrying objects.
Parameters that were objectively assessed by the
treating physician on a 1–10 pain scale (1 = no pain;
* GELITA Health GmbH, Eberbach, Germany
1486 24-week study of collagen hydrolysate
© 2008 Informa UK Ltd – Curr Med Res Opin 2008; 24(5)
10 = severe pain) were also found to have similarly
improved after 12 weeks.
The lack of a control group in this study limits
the use of these data and prevents the conclusion
being drawn that collagen hydrolysate is suitable as
an agent for primary prevention purposes. However,
the findings of this study suggested the importance
of conducting another similar but well-controlled
study. The aim of the current study is to extend these
earlier findings with a more rigorous study design (that
includes a control group) to determine if consuming a
standardized preparation of collagen hydrolysate can
reduce activity-related joint pain in athletes who do
not have a diagnosis of OA.
Methods
Study design and population
This was a single-center, prospective, randomized,
double-blind, placebo-controlled trial.
The original target was chosen to be 150 participants.
Between September 2005 and June 2006, 147
individuals were recruited who were active as student
athletes either on a varsity team or a club sport who
complained about joint pain or joint discomfort due to
joint stress, injury, surgical outcome, or trauma.
Potential subjects were excluded from the study if
they met the following exclusion criteria:
●
●
●
●
They did not have joint pain or joint discomfort
They had an acute injury of a joint or an inflammatory process
They ingested glucosamine, chondroitin, or other
nutritional supplements that may be indicated for
treatment of joint pain and OA
People who from a clinical perspective were
considered likely to increase their dose of analgesic
medication during the 24-week study phase of the
trial because of severe symptoms of arthralgia
The study was approved by the university IRB and
students willing to participate signed informed
consent.
According to the study schedule, after signing
informed consent forms and randomization, the
subjects were seen five times: for a baseline visit and
four follow-up visits. Thus, the fifth visit corresponded
with the subjects’ final evaluation. The study schedule
contained a 6-week interval between two consecutive
visits.
The randomization list was drawn up using the
nQuery Advisor computer program version 5.0
(Statcon, Witzenhausen, Germany). Only one person,
author K.F., had the codes for the randomization. The
© 2008 Informa UK Ltd – Curr Med Res Opin 2008; 24(5)
study partici­pants were assigned to receive either 10 g
of collagen hydrolysate per day in the form of a vial
containing 25 mL of a liquid formulation (n = 73) or a
placebo that consisted of 25 mL of a liquid formulation
containing xanthan (n = 74). The vials containing both
the nutritional supplement and the placebo were indis­
tinguishable from each other in terms of color, taste, or
viscosity. A representative of the sponsor was the only
person able to break the code.
During the baseline visit, a history was taken in
relation to musculoskeletal diseases and subjects were
given a physical exam. Questions were asked about
the presence of pain in the knee, hip, shoulder, ankle,
wrist, elbow, neck, and back. Apart from the location
of the pain, data were collected concerning the cause
of any joint pain, such as degenerative disease, sports
injury, joint deformity, or genetic predisposition, and
also the time span during which the subjects already
had experienced pain in those respective joints (< 1
year, 1–5 years, 6–10 years or > 10 years).
A medical history was taken. Consumption of pain
relievers, anti-inflammatory agents, cyclo-oxygenase II
inhibitors, and other over-the-counter analgesics was
recorded and information was collected in relation to
alternative therapies like acupuncture, hydrotherapy,
electric stimulation, massage therapy, joint condition­
ing/training, or topical application of ice or heat.
At the baseline visit, joint discomfort was recorded.
Subjects’ complaints of joint discomfort were recorded
using a visual analogue scale (VAS), which was divided
into increments of 1 to 10 in relation to the severity
of symptoms, with 1 = no manifestation of symptoms
and 10 = very severe symptoms. The physician rated
the following parameters: joint pain at rest, joint pain
related to exertion, restricted ability to move, and state
of inflammation.
Study participants rated their subjective symptoms
using the same VAS scale. The degree of pain was selfrated, such as pain when walking, standing, running a
straight line, running and changing direction, carrying
objects, lifting, extending arms, rotating the shoulder,
reaching and throwing, and at rest. The first five
parameters were assessed when joints pertaining to the
lower extremities were affected and the symptoms of
the remaining parameters were recorded when joints
of the upper extremities were affected.
When study participants presented for the following
four visits, both the treating physician and the patients
documented the objective and subjective findings as
described for the initial visit. At each visit, patients were
asked to provide information on consumption of pain
relievers, anti-inflammatory agents, cyclo-oxygenase II
inhibitors, and other over-the-counter analgesics.
Inform­ation was collected on alternative therapies,
including acupuncture, hydrotherapy, electric
24-week study of collagen hydrolysate Clark et al. 1487
stimul­a tion, massage therapy, joint conditioning/
training, or topical applica­t ion of ice or heat. The
fourth follow-up visit (the fifth visit in the study) was
considered the final visit and the end of the study. At
baseline and each visit, subjects were given either the
study treatment (collagen hydrolys­ate) or the placebo at
baseline. They received either once-monthly or twicemonthly packages, depending on how many vials they
had remaining from the previous visit. Compliance of
medication was determined by asking participants to
return all unused vials.
The primary end points of the study were defined
as the comparison of the numerical differences of the
scores of the VAS between both groups (collagen
hydrolysate vs. placebo) after 24 weeks of treatment.
These scores were calculated by subtracting the score
of a particular parameter of visit 1 from the score
recorded during visit 5. This method was applied to
both the parameters assessed by the physicians and by
the study participants.
second smallest p-value was compared with the number
α/14, and so on, up to the point where the highest
p‑value was compared with the number 0.05, which is
α/1. According to this theory, all the p-values that turn
out to be smaller than the comparative number can
be regarded as representing a statistically significant
difference between the groups. The confirmatory
analysis provided data in terms of statistical confirmed
superiority or lack of statistical confirmed superiority
for one of the treatment groups.
Explorative analysis
An explorative analysis was used to evaluate secondary
end points. The χ2-test was chosen to determine whether
there was a difference between both treatment groups.
However, multiple testing was not performed for the
secondary end points. The results of the explorative
analysis are of a descriptive nature and do not provide
any conclusion as to whether one treatment group is
superior over the other group.
Descriptive analysis
The parameters that were evaluated for both treatment
groups were analyzed in a descriptive manner.
The results were listed in tables, with nominal and
categorical data indicated by absolute and relative
frequency. Continuous data were listed as mean,
standard error of the mean, median, standard deviation,
variance, range, minimum and maximum.
Confirmatory analysis
A confirmatory analysis was carried out for the
parameters that had been defined as primary end
points. The scores that reflected the level of joint
discomfort on the VAS were considered as continuous
data. The test that was used to determine whether there
was a difference between the two treatment groups
was the non-parametric Mann–Whitney U-test. The
significance level was defined to be α = 0.05. A twosided test was performed at all times to substantiate
superiority of one group over the other. Because no
hierarchy between the variables had been determined
prior to the study, multiple testing was performed by
adapting the significance levels according to the theory
of Bonferroni–Holm.
For 15 separate tests, which were due to 15 separate
primary end points, the global significance level of
0.05 had to be adapted to 15 local significance levels.
According to Bonferroni–Holm, the calculated p-values
appeared in rising order, the smallest p-value being
listed first and the highest p-value being listed last.
The smallest p-value was subsequently compared with
the number 0.0033 (which is equivalent to α/15), the
1488 24-week study of collagen hydrolysate
Results
In all, 147 individuals were recruited and gave informed
consent to participate in the trial (Figure 1). Ten
subjects were considered ineligible when they appeared
for their first visit based on the treating physician’s or
their own judgment because their joint discomfort had
dis­appeared. Seventeen subjects did not present for the
first visit and all the other visits and were lost to followup. Eight more participants were excluded from data
management because of faulty documentation during
the first visit. Four participants suffered an adverse
event: two subjects suffered a new joint injury, one
subject had to undergo surgery, and one subject was
the victim of a car accident.
Eleven subjects presented fewer than three times
for the scheduled visits, thus rendering a meaningful
evalua­t ion of their data impossible. In summary,
97 patients could be considered as the analysis
population (46 (23 male, 23 female) receiving
collagen hydrolysate and 51 (22 male, 29 female)
receiving placebo). Twenty-five subjects out of these
97 participants presented fewer than five times for
the scheduled visits, which conse­quently results in a
per-protocol (PP) population of 72 individuals.
The baseline age, height, weight, and BMI of the
study population showed no statistically significant
differences between the treatment and placebo groups,
as shown in Table 1. In addition, no difference between
treatment group and placebo group could be observed
in vitamin and mineral supplement use, although it
approached significance ( p = 0.062, data not shown).
© 2008 Informa UK Ltd – Curr Med Res Opin 2008; 24(5)
The distrib­ution of joint discomfort among various
joints is shown in Table 2. The duration of pain was
similar in both groups ( p = 0.565, data not shown).
In 13 out of 97 subjects (13.4% of the analysis
popula­tion), only one joint was affected by pain. The
remainder of study participants had joint discomfort
in two to eight joints. Each individual in the analysis
population averaged arthralgia in 3.01 joints. For each
joint that was affected by pain, the investigators used a
separate sheet to collect data on the joint, so they were
able to follow up on individual joints for each affected
joint during the 24-week study phase.
The occurrence of pain in various joints was wellbalanced between treatment groups (paired t-test for
Subjects being randomized
(n = 147)
Number of individuals being
documented
(n = 120)
n = 27 (no pain or all visits
missing)
n = 8 (data not plausible)
ITT-population: Individuals
having presented for the
baseline-visit
(n = 112)
Population without AE
n =108
n = 4 (adverse events)
n = 11 (< 3 visits)
Analysis population
(n = 97)
n = 25 (≥ 3 visits < 5)
Per-protocol population
(n = 72)
Figure 1. Disposition of subjects in the trial
equivalence). The joint that most participants (63
subjects: 29 treatment, 34 placebo) indicated for joint
discomfort was the knee.
After assessment at baseline, no significant difference
between the treatment and placebo group was found
for cause of pain, i.e., degenerative disease, sports
injury, joint deformation, or genetic predisposition.
No differ­ence between groups was found ( p = 0.565)
for length of time for pain. Furthermore, when intake
of medications (anti-inflammatory drugs, COX-II
inhibitors, and other pain relievers) was analyzed,
15 out of 44 collagen hydrolysate patients, and 17 of
the 49 placebo patients reported taking some form
of medication. Therefore, no statistically significant
difference between the treatment and placebo groups
existed ( p = 0.951).
The baseline severity of symptoms using a VAS
ranging from 1 to 10 are shown in Table 3. Based on
these results, the authors concluded that for the baseline
visit, both groups were equal in severity of symptoms
as assessed by the treating physicians, except for the
parameter ‘restricted ability to move.’ As far as that
particular parameter was concerned, the symptoms
were judged as being less severe for the group receiving
the nutritional supplement (collagen hydrolysate:
1.36 ± 1.64; placebo: 1.67 ± 1.74, p = 0.038).
In compliance with the study protocol, at baseline
and at each visit participants rated the severity of
various symptoms using a VAS that ranged from 1 to
10. The group of participants receiving the supplement
was equal to the group receiving the placebo in
terms of the severity of symptoms as judged by study
participants (Table 4), except for the parameter joint
pain when rotating shoulder ( p = 0.018). This finding
suggested that subjects receiving placebo had more
severe symptoms than individuals receiving collagen
hydrolysate in this one parameter at baseline.
Primary efficacy outcomes
Interestingly, differences between groups became
apparent at 24 weeks; those differences reached
statistical significance for a number of parameters.
Table 1. Baseline characteristics of the study population
All patients
n = 97
Treatment
Placebo
p-value
20.1 ± 1.47
19.9 ± 1.50
20.3 ± 1.43
0.179
45/52
23/23
22/29
0.499
Height, m
1.80 ± 0.24
1.83 ± 0.32
1.77 ± 0.14
0.254
Weight, kg
76.7 ± 17.8
77.9 ± 16.2
75.6 ± 19.2
0.530
BMI, kg/m2
23.9 ± 4.19
23.8 ± 4.36
23.9 ± 4.08
0.901
Age, years
Male/female
© 2008 Informa UK Ltd – Curr Med Res Opin 2008; 24(5)
24-week study of collagen hydrolysate Clark et al. 1489
Table 2. Distribution of joint discomfort among various joints
Knee pain:
both knees
Collagen hydrolysate
Placebo
17
20
Hip pain:
both hips
Collagen hydrolysate
Placebo
3
6
Shoulder pain:
both shoulders
Collagen hydrolysate
Placebo
8
5
Ankle pain:
both ankles
Collagen hydrolysate
Placebo
7
10
Wrist pain: both
wrists
Collagen hydrolysate
Placebo
2
2
Elbow pain:
both elbows
Collagen hydrolysate
Placebo
2
0
Neck pain:
no
Collagen hydrolysate
Placebo
28
28
Knee pain:
left knee
3
9
Hip pain:
left hip
5
0
Shoulder pain:
left shoulder
2
5
Ankle pain:
left ankle
6
2
Wrist pain: left
wrist
4
5
Elbow pain:
left elbow
1
3
Neck pain:
yes
8
7
Back pain:
no
Back pain:
yes
21
20
17
17
Collagen hydrolysate
Placebo
Knee pain:
right knee
9
5
Hip pain:
right hip
2
1
No
knee pain
p-value χ2-test
15
11
0.172
No
hip pain
33
34
Shoulder pain:
right shoulder
No
shoulder pain
10
13
23
19
Ankle pain:
right ankle
No
ankle pain
5
4
Wrist pain: right
wrist
9
3
Elbow pain:
right elbow
5
7
Neck pain:
not sure
4
4
Back pain:
I’m not sure
2
4
25
28
0.098
0.434
0.424
No
wrist pain
26
30
0.336
No
elbow pain
33
29
0.315
Neck pain:
sometimes
1
0
0.797
Back pain:
sometimes
1
0
0.639
Table 3. Severity of symptoms of patients (n = 97) at baseline, as rated by physicians
Physician-rated symptoms
All subjects
Treatment subjects
Placebo subjects
p-value
Joint pain at rest
2.12 ± 1.88
2.36 ± 2.14
1.87 ± 1.53
0.200
Joint pain related to exertion
5.36 ± 1.81
5.34 ± 1.88
5.39 ± 1.75
0.760
Restricted ability to move
1.52 ± 1.69
1.36 ± 1.64
1.67 ± 1.74
0.038*
State of inflammation
1.90 ± 1.52
2.03 ± 1.81
1.76 ± 1.16
0.675
*Indicates between-group difference < 0.05
For the parameter joint pain at rest, which was based
on the physicians’ judgment, collagen hydrolysate
was statistically significantly superior to placebo ( p =
0.025 (CH: –1.37 ± 1.78, placebo: –0.90 ± 1.74)).
For the parameters walking (Figure 2), standing at
rest (Figure 3), carrying objects, and lifting, which
were all based on the study participants’ judgment,
collagen hydrolysate was statistically significantly
1490 24-week study of collagen hydrolysate
superior compared with placebo, with p-values of
0.007 (CH: –1.11 ± 1.98, placebo: –0.46 ± 1.63 ),
0.011 (CH: –0.97 ± 1.92, placebo: –0.43 ± 1.74),
0.039 (CH: –0.81 ± 1.77, placebo: –0.39 ± 1.56),
0.014 (CH: –1.45 ± 2.11, placebo: –0.83 ± 1.71) and
0.018 (CH: –1.79 ± 2.11, placebo: –1.26 ± 2.09),
respectively. As the study participants all improved
(that is, they experienced decreased pain) during
© 2008 Informa UK Ltd – Curr Med Res Opin 2008; 24(5)
Table 4. Severity of joint pain, as rated by patients at baseline. *Indicates between-group difference < 0.05
Joint pain type
All subjects
Treatment
Placebo
p-value
While walking
2.15 ± 1.97
2.37 ± 2.26
1.93 ± 1.61
0.562
When standing
2.04 ± 2.00
2.21 ± 2.16
1.87 ± 1.8
0.563
At rest
1.77 ± 1.80
1.92 ± 2.06
1.63 ± 1.48
0.893
When running a straight line
2.86 ± 2.13
2.93 ± 2.41
2.79 ± 1.82
0.769
When running and changing direction
3.43 ± 2.39
3.48 ± 2.58
3.38 ± 2.18
0.994
When carrying objects
2.88 ± 2.09
3.07 ± 2.32
2.69 ± 1.82
0.275
When lifting
3.69 ± 2.19
3.97 ± 2.38
3.42 ± 1.97
0.071
When extending arms
1.52 ± 1.85
1.51 ± 2.03
1.53 ± 1.66
0.094
When rotating the shoulder
1.54 ± 1.87
1.45 ± 1.99
1.64 ± 1.76
0.018*
When reaching
1.45 ± 1.64
1.44 ± 1.81
1.45 ± 1.47
0.141
When throwing
2.09 ± 2.23
2.03 ± 2.34
2.16 ± 2.13
0.164
Figure 2. Change of pain perception for the parameter
walking in the analysis population (n = 97) according to the
visual analogue scale throughout the study phase (difference:
score visit 5 – score visit 1) at 24 weeks. The numbers
illustrated are the mean ± the 95% confidence interval. The
larger the number, the greater the decrease in pain
Figure 3. Change of pain perception for the parameter
standing in the analysis population (n = 97) as recorded with
the use of the visual analogue scale during the study (difference:
score visit 5 – score visit 1) at 24 weeks. The numbers
illustrated are the mean ± the 95% confidence interval. The
larger the number, the greater the decrease in pain
the trial, regardless of whether they were part of the
nutritional supplement or the placebo group, the
absolute figures of the differences are always negative.
So, the more negative a result, the more pronounced
is the clinical improvement. These results imply that
genuine differences in the evolution of pain scores
were seen when the effect of collagen hydrolysate on
joint discomfort was compared with placebo.
Next, we hypothesized that, by analyzing the scores
of joint discomfort to one joint, the knee, we might find
more clinically meaningful results. From the perspective
of rheumatological research, a separate analysis
of collagen hydrolysate’s effect on knee arthralgia
appeared meaningful, as most clinical trials exploring
the disease-modifying actions of investigational drugs
focus on the knee.
© 2008 Informa UK Ltd – Curr Med Res Opin 2008; 24(5)
24-week study of collagen hydrolysate Clark et al. 1491
Data pertaining to 96 individuals were used for this
post hoc subgroup analysis. In 18 subjects, there were
data missing regarding all five scheduled visits. As an
adverse event occurred in three individuals, those
data had to be excluded from the analysis as well. In
four study participants, data regarding the first visit
were missing, so those subjects were also excluded
from the statistical analysis, and eight individuals
presented for less than three visits. Therefore, the
final analysis subpopula­t ion was equivalent to 63
subjects. Among those 63 individuals, 18 subjects
presented either for three or four visits, which
subsequently corresponds to a PP population of 45
study participants (Figure 4).
The 63 subjects whose data were used for a subgroup
analysis had an average age of 20.1 years, an average
height of 1.77 m, an average weight of 75.1 kg, and an
average body mass index (BMI) of 23.8 kg/m2. When
the anthropometric data of the treatment group (n =
29) and the placebo group (n = 34) were compared
by using a t-test, a difference in age could be noted
between the groups (19.7 ± 1.50 years vs. 20.6 ± 1.42
years ( p = 0.017)), which from the clinical point of
view was irrelevant. Both groups were equal in relation
to their use of medication ( p = 0.768) and their use of
alternative therapies ( p = 0.458).
For the subgroup analysis, the parameters assessed by
the physicians remained the same, as shown in Table 5,
except for the parameter joint pain at rest ( p = 0.028).
This finding suggests that, except for the severity of
this particular symptom, which was more severe for
the participants in the collagen hydrolysate group, both
groups were equal at the baseline visit in terms of the
parameters that were assessed by the physicians.
From the perspective of the subjects, only five out
of the 11 parameters recorded on the clinical report
form were assessed for knee arthralgia, as those
parameters focused on the severity of symptoms
when the lower extremities were actively used.
Subjects being recruited
(n = 96)
n = 18 (all visits missing)
Documented population
(n = 78)
n = 4 (visit 1 missing)
Subjects having presented
for the baseline visit
(n = 74)
n = 3 (adverse events)
Population without AE
(n = 71)
n = 8 (< 3 visits)
Analysis subpopulation
(n = 63)
n = 18 (≥ 3 visits < 5)
Per-protocol population
(n = 45)
Figure 4. Disposition of subjects (subgroup analysis (knee
arthralgia)) in the trial
Table 5. Patient parameters assessed by physicians, knee arthralgia subgroup analysis (n = 63)
Joint pain type
All subjects
Treatment
Placebo
p-value
At rest
2.28 ± 1.94
2.67 ± 2.23
1.91 ± 1.55
0.028
Related to exertion
5.41 ± 1.77
5.49 ± 1.82
5.34 ± 1.73
0.588
Restricted ability to move
1.44 ± 1.57
1.36 ± 1.56
1.52 ± 1.59
0.229
State of inflammation
1.98 ± 1.54
2.17 ± 1.80
1.80 ± 1.23
0.269
Table 6. Subjects’ self-report of symptoms of the lower extremities, subgroup analysis (n = 63)
Joint pain type
All subjects
Treatment
Placebo
p-value
When walking
2.32 ± 2.10
2.64 ± 2.45
2.02 ± 1.67
0.333
When standing
2.18 ± 2.07
2.43 ± 2.34
1.95 ± 1.77
0.565
At rest
1.96 ± 1.91
2.19 ± 2.22
1.74 ± 1.55
0.489
When running a straight line
3.00 ± 2.19
3.29 ± 2.60
2.74 ± 1.69
0.293
When running and changing direction
3.72 ± 2.42
4.02 ± 2.68
3.43 ± 2.11
0.096
1492 24-week study of collagen hydrolysate
© 2008 Informa UK Ltd – Curr Med Res Opin 2008; 24(5)
Secondary efficacy outcomes
Those parameters were joint pain when walking,
joint pain when standing, joint pain at rest (subject’s
perspective), joint pain when running a straight line,
and joint pain when running and changing direction.
These parameters are shown in Table 6. Apart from
one parameter as judged by the physicians, the
severity of symptoms was equivalent in both the
group receiving the nutritional supplement and in the
group being treated with the placebo.
When the primary end point was calculated on the
basis of the analysis population for knee arthralgia, a
statistically significant difference became apparent
for the following parameters: joint pain at rest, with
a p = 0.001 (CH: –1.67 ± 1.89, placebo: –0.86 ± 1.77)
(Figure 5); joint pain when walking, with a p = 0.003
(CH: –1.38 ± 2.12, placebo: –0.54 ± 1.65) (Figure 6);
joint pain when standing, with a p = 0.015 (CH:
–1.17 ± 2.06, placebo: –0.50 ± 1.68); joint pain at rest
(subject), with a p = 0.021 (CH: –1.01 ± 1.92, placebo:
–0.47 ± 1.63); joint pain when running a straight line,
with a p = 0.027 (CH: –1.50 ± 1.97, placebo: –0.80 ±
1.66); and joint pain when running and changing
direction, with a p = 0.026 (CH: –1.87 ± 2.18, placebo:
–1.20 ± 2.10).
When the intake of medication (anti-inflammatory
drugs, COX-II inhibitors, and other pain relievers)
was analyzed throughout the study for the analysis
population (n = 97) at the final evaluation (visit 5),
no difference between groups was seen ( p = 0.301);
however, when alternative therapies were analyzed, a
statistically significant difference was noted between
the collagen hydrolysate and the placebo group ( p <
0.001). Specifically, at visit 5, the group taking collagen
hydro­lysate reported using alternative therapies 12
times, while subjects taking placebo reported using
alternative therapies 39 times (Table 7).
The same results were found for the subgroup analysis
(n = 63 subjects). No difference was noted in terms of
the medication that was taken by the subjects at visit 5
( p = 0.227). However, a statistically significant differ­
ence was found between both groups when the use of
alternative therapies was analyzed during the final visit
( p < 0.001), with the collagen hydrolysate group report­
ing the use of alternative therapies 15 times compared
with 30 times by the placebo group. This difference was
only observed at visit 5 and not at the other four visits.
Figure 5. Change of pain perception for the parameter joint
pain at rest (physicians’ judgment) for knee arthralgia in
subgroup analysis population (n = 63) as recorded with the
use of the visual analogue scale during the study (difference:
score (visit 5) – score (visit 1)). The numbers are the
mean ± the 95% confidence interval. The larger the number,
the greater the decrease in pain
Figure 6. Change of pain perception for the parameter joint
pain when walking for knee arthralgia in subgroup analysis
population (n = 63) as recorded with the use of the visual
analogue scale during the study (difference: score (visit 5) –
score (visit 1)). The numbers illustrated are the mean ± the
95% confidence interval. The larger the number, the greater
the decrease in pain
© 2008 Informa UK Ltd – Curr Med Res Opin 2008; 24(5)
24-week study of collagen hydrolysate Clark et al. 1493
Table 7. Number of subjects (collagen hydrolysate or placebo groups) reporting they used alternative therapies (e.g., acupuncture,
hydrotherapy, electric stimulation, massage therapy, joint conditioning/training, or topical application of ice or heat) at visit 5
Patient group
n
Collagen hydrolysate
Placebo
p-value
Analysis population
97
12
39
< 0.001
Subgroup analysis population
63
15
30
< 0.001
Adverse events
During the study, four adverse events were recorded.
Two subjects (one CH and one placebo) suffered a
new joint injury, one subject (placebo) had to undergo
surgery and one subject (CH) had a car accident. None
of the adverse events were determined to be treatment
related. Since these adverse events interfered with an
assessment of the study participants’ joint condition,
the data pertaining to those subjects were excluded
from statistical analysis.
Discussion
When taken orally by patients diagnosed with OA,
collagen hydrolysate is a nutritional supplement that
results in an increase of mobility, a decrease of pain, and
a reduction in the dependency on analgesics5,6,9,12. Studies
of this supplement are available in the medical literature
with more than 2000 patients who had experienced
degenerative joint disease (reviewed in Bello and
Oesser13). However, many of these studies were openlabel and/or did not provide statistical information in
published reports. Also, the studies may be summarized as
secondary prevention efforts, because they offer evidence
suggesting the efficacy of the nutritional supplement in
patients who had been diagnosed with OA.
The observational study that was performed by
Flechsenhar and Alf12 with athletes from the Olympic
Games site at Essen, Germany, was the first clinical
study that investigated the use of collagen hydrolysate
as a nutritional supplement to reduce symptoms of
joint damage, with the hope that this change would
reflect improvements in joint health. In that study,
individuals were recruited who had not been diagnosed
with degenerative joint disease but who complained
about joint pain that both the treating physician and
the subjects interpreted as being a result of stressful
exercising. It was reported that 78% of individuals at
the end of the study noticed substantial improvement
of their joint symptoms. That finding could be
numerically substantiated as the study participants at
the Olympic Games site in Essen improved by up to
five points on the visual analogue scale.
In spite of the fact that the results of the German
observational study confirmed findings of previous
well-controlled trials, the results of that study had the
1494 24-week study of collagen hydrolysate
drawback that it did not include a control group. In
rheumatology research, placebo effects and interpret­
ations of efficacy of disease-modifying agents are closely
intertwined and are sometimes hard to distinguish
from one another. Consequently, the findings from the
German observational study gave rise to the need for
a placebo-controlled trial with athletes experiencing
exercise-related joint pain. The result was the study
described in this report.
The objective of this prospective clinical trial was to
determine whether the administration of 10 g/day of
collagen hydrolysate for 24 weeks, i.e. the same dosage and
the same timespan that was defined in many OA trials,
would translate into improvement of joint symptoms in
individuals who did not have degenerative joint disease
but who were physically active. Thus, recruiting young,
healthy, and active individuals to participate in this
research project was a crucial aspect of the trial.
From the 147 individuals who started the study,
97 subjects could be included in the analysis, which
represents 66% of participants. From the perspective
of a nutritional supplement study, a dropout rate of
34% is reasonable, underlining the willingness and the
proactive attitude of the study participants to comply
with the protocol.
Anthropometric data and information related to
previous or present treatment modalities and the cause
of pain revealed that the baseline characteristics of both
the treatment and the placebo group were equivalent.
The same was true when the subgroup of individuals
complaining of arthralgia of the knee was submitted to
an additional analysis.
When individuals with arthralgia of various joints
were assessed during the baseline visit (visit 1) for the 15
parameters that were recorded on the visual analogue
scales, apart from one parameter (pain when rotating
shoulder), no difference of severity of symptoms was
noted between the groups, which meant that a change
in parameters during the 24-week study period would
reflect an effect of the nutritional supplement collagen
hydrolysate on joint symptoms.
The end point that yielded the most interest was
the difference in scores for the analysis population (n =
97), when the findings of the final follow-up visit were
compared with the status that was recorded during
visit 1. Among the parameters that were evaluated, the
parameter pain at rest as assessed by the physician (CH
© 2008 Informa UK Ltd – Curr Med Res Opin 2008; 24(5)
vs. placebo (–1.37 ± 1.78 vs. –0.90 ± 1.74 ( p = 0.025))
and the following five parameters as assessed by the study
participants: joint pain when walking (–1.11 ± 1.98 vs.
–0.46 ± 1.63 ( p = 0.007)), joint pain when standing
(–0.97 ± 1.92 vs. –0.43 ± 1.74 ( p = 0.011)), joint pain
at rest (–0.81 ± 1.77 vs. –0.39 ± 1.56 ( p = 0.039)), joint
pain when carrying objects (–1.45 ± 2.11 vs. –0.83 ± 1.71
( p = 0.014)) and joint pain when lifting (–1.79 ± 2.11 vs.
–1.26 ± 2.09 ( p = 0.018)) were found to be statistically
significant. Thus, in each of these parameters, the subjects
taking collagen hydrolysate reported less joint discomfort
than subjects taking placebo.
When the use of alternative therapies was submitted
to an analysis, it became apparent that there was a
significant difference between the collagen hydrolysate
and the placebo groups. As shown in Table 7, for the
analysis population (n = 97), the collagen hydrolysate
group reported using alternative therapies 12 times,
while the placebo patients reported using alternative
therapies 39 times ( p < 0.001) during the final followup visit (visit 5). In the prior visits, no difference
regarding alternative therapies was noted. These findings
can be interpreted to indicate that it required more
than 3 months for the effect of collagen hydrolysate
to become clinically relevant from the point of view
of a primary prevention approach. It appeared that the
individuals who had been assigned to the placebo group
were possibly lacking some beneficial effect that the
subjects assigned to the nutritional supplement group
perceived. The number of alternative therapies through­
out the study decreased in the nutritional supple­ment
group whereas it increased in the placebo group, thus
graphically resembling two curves that diverted.
Because 65.3% of individuals initially recruited for
the study (96 participants out of 147 who had signed
the informed consent form), which corresponds to
64.9% on the analysis population (63 subjects out
of 96), complained of arthralgia of the knee, and
because most studies in rheumatology are exclusively
standardized for the knee, a post hoc analysis for
that particular joint was performed. The analysis of
anthropometric data and the description of previous
treatment modalities showed equivalence between
the nutritional supplement and the placebo group
and indicated that an analysis of that subgroup would
reveal meaningful results. There was a difference in age
between both groups (CH vs. placebo (19.7 vs. 20.6
years ( p = 0.017)); however, from the clinical point of
view, this difference would not be expected to produce
any discrepancies in the study data.
When the primary end points for the analysis
subpopulation (63 subjects) were considered, six out of
nine parameters were statistically significant. Again, it
was the parameter joint pain at rest that was assessed by
the physician with a p = 0.001 (–1.67 ± 1.89 vs. –0.86 ±
© 2008 Informa UK Ltd – Curr Med Res Opin 2008; 24(5)
1.77), plus the remaining five parameters that were based
on the participants’ assessments, such as joint pain when
walking with a p = 0.003 (–1.38 ± 2.12 vs. –0.54 ± 1.65),
joint pain when standing, with a p = 0.015 (–1.17 ± 2.06
vs. –0.50 ± 1.68), joint pain at rest with a p = 0.021
(–1.01 ±1.92 vs. –0.47 ± 1.63), joint pain when running
a straight line, with a p = 0.027 (–1.50 ± 1.97 vs. –0.80 ±
1.66) and joint pain when changing direction with a p =
0.026 (–1.87 ± 2.18 vs. –1.20 ± 2.10).
The use of alternative therapies also appeared to be
important when considered as a secondary outcome
parameter. During the final follow-up visit (visit 5),
participants assigned to the placebo group had a clear
preference for using this therapeutic modality (30 times
for the placebo patients compared with 15 times for the
collagen hydrolysate patients, p < 0.001) (Table 7).
These results indicated that administration of
10 g/day of collagen hydrolysate for 24 weeks translated
into improvement of joint symptoms in individuals
who did not have degenerative joint disease but who
were physically active.
The findings of this study were analogous to the
results that were shown in classical OA studies with
collagen hydrolysate. In osteoarthritis, according to the
results of previous studies, collagen hydrolysate starts
to exert a beneficial effect on joint symptoms after
a period of roughly 3 months. Approximately 80%
of patients start perceiving an improvement of their
joint symptoms after 10 or 12 weeks. In this study, a
difference between the two groups did not become
visible after 12 weeks; however, it became clearly
apparent after 24 weeks. This also became evident
when the secondary parameter alternative therapies
was assessed. It was at the very end of the study that
the placebo group increasingly used this modality.
So, when the action of collagen hydrolysate on joint
disorders is considered, a certain period of time must
elapse before an effect manifests.
The parameter effect size (ES) is an important tool in
research in rheumatology for assessing the magnitude
of a therapeutic effect14. It is defined according to the
following formula:
ES =
∆ Active agent − ∆ Placebo
Pooled standard deviation
which means that the effect size represents a ratio
that is calculated by determining the difference of
the changes of a particular parameter exerted by
the drug and the placebo, divided by the arithmetic
mean of both standard deviations. If the effect size
for a particular agent as determined by a parameter
that may serve as the end point in a clinical trial is
lower than 0.2, then the effect exerted by that agent
24-week study of collagen hydrolysate Clark et al. 1495
is considered weak. If it is higher than 0.8, it is
considered strong.
When the effect size for the parameter joint pain
when walking on the basis of the analysis population
for all the joints (n = 97) is calculated, the result
corresponds to 0.36.
When the effect size for the same parameter is
calculated on the basis of the analysis subpopulation
for knee arthralgia (n = 63), then the result is 0.45.
One should take into consideration that the subjects
recruited for this study were healthy, young individuals.
Consequently, the effect size that has been shown in
this group of individuals may give a hint in regard to
the potential that collagen hydrolysate has in view of a
primary-prevention approach.
A limitation of the study was the lack of standard­
ization in relation to sports activities. The investigators
included several types of sports and did not randomize
individuals by the activities they participated in. It
may have been better to include only athletes involved
in similar sports (e.g., football or basketball players);
however, for practical reasons, this was not feasible.
Conclusion
The purpose of this study was to determine whether
administration of 10 g/day of collagen hydrolysate
would reduce joint pain in individuals with joint pain
due to strenuous effort and physical exercise. For this
research project, athletes of Penn State University were
chosen as study participants.
The design of the study was appropriate to reveal
that collagen hydrolysate as a nutritional supplement
ingested over a period of 24 weeks by physically active
young adults was efficacious in reducing symptoms of
joint discomfort.
The results of this trial provide data supporting the
view that collagen hydrolysate as a nutritional supple­
ment may be administered to athletes to reduce
the symptoms of joint pain associated with athletic
activity. Taken together with preclinical studies which
suggest that oral collagen hydrolysate reaches joints and
stimulates joint tissues, athletes consuming collagen
hydrolysate can potentially improve their joint health
and reduce pain symptoms associated with strenuous
athletic activity. Further research will clarify additional
benefits from collagen hydrolysate.
Acknowledgment
Declaration of interest: This research was sponsored
by GELITA Health GmbH, Eberbach, Germany. The
authors acknowl­edge the contributions of Dr. Imma U.
Fischer, (Institute for Biostatistics, Tübingen, Germany)
for the statistical analysis of the trial data. Editorial
assistance was provided by Oliver Fultz, ACCESS
Medical Group, Chicago, IL, USA. The nutritional
supplement collagen hydrolysate used for this research
project was provided by GELITA Health GmbH,
Eberbach, Germany.
References
1. Oesser S, Adam M, Babel W, Seifert J. Oral administration
of (14)C labeled gelatin hydrolysate leads to an accumulation
of radioactivity in cartilage of mice (C57/BL). J Nutr 1999;
129:1891-5
2. Oesser S, Seifert J. Stimulation of type II collagen biosynthesis
and secretion in bovine chondrocytes cultured with degraded
collagen. Cell Tissue Res 2003;311:393-9
3. Oesser S, Seifert J. Impact of collagen fragments on the
synthesis and degradation of the extracellular matrix (ECM) of
cartilage tissue. Orthopaedische Praxis 2005:565-8
4. Ng KW, Saliman JD, Tan K, Statman LY, et al. Collagen
hydrolysate increases the mechanical properties and type II
collagen synthesis of tissue engineered articular cartilage. Paper
presented at: 51st Annual Meeting of the Orthopedic Research
Society; February 2005, Washington, DC
5. Krug E. Zur unterstützenden Therapie bei Osteo- und Chondro­
pathien [On supportive therapy for osteo- and chondropathy].
Z Erfahrungsheikunde 1979;11:930-8
6. Götz B. Gut genährter Knorpel knirscht nicht mehr [Wellnourished cartilage does not grind]. Ärztl Prax 1982;92:
3130-4
7. Oberschelp U. Individuelle Arthrosetherapie ist möglich
[Individual arthrosis therapy is possible]. Therapiewoche 1985;44:
5094-7
8. Seeligmüller K. Kann eine Gelatine/L-Cystin-Mischung die
Kollagen- und Proteoglykansynthese stimulieren? Therapiewoche
1989;39:3153-7
9. Adam M. Welche Wirkung haben Gelatinepräparate? Therapie
der Osteoarthrose [What effects do gelatin preparations have?].
Therapiewoche 1991;41:2456-61
10. Moskowitz RW. Role of collagen hydrolysate in bone and joint
disease. Semin Arthritis Rheum 2000;30:87-99
11. Zuckley L, Angelopoulou K, Carpenter MR, et al. Collagen
hydrolysate improves joint function in adults with mild
symptoms of osteoarthritis of the knee. Med Sci Sports Exerc
2004;36(Suppl):S153-S4
12. Flechsenhar K, Alf D. Ergebnisse einer Anwendungsbeobachtung
zu Kollagen-Hydrolysat CH-Alpha. Orthopaedische Praxis
2005;9:486-94
13. Bello AE, Oesser S. Collagen hydrolysate for the treatment
of osteoarthritis and other joint disorders: a review of the
literature. Curr Med Res Opin 2006;22:2221-32
14. Hochberg MC. What a difference a year makes: reflections on
the ACR recommendations for the medical management of
osteoarthritis. Curr Rheumatol Rep 2001;3:473-8
CrossRef links are available in the online published version of this article:
http://www.cmrojournal.com
Paper CMRO-4309_4, 10:22-21.04.08
Accepted for publication: 29 February 2008
Published Online: 15 April 2008
doi:10.1185/030079908X291967
1496 24-week study of collagen hydrolysate
© 2008 Informa UK Ltd – Curr Med Res Opin 2008; 24(5)